Working memory (WM) refers to the active retention of information when it is not accessible in the environment. It is of central interest for understanding human behavior and health, because it is predictive of other cognitive factors and is compromised in many neurological and psychiatric conditions. However, the neural bases underlying WM are not well understood. Interestingly, WM and attention (a related construct) seem to involve overlapping brain areas (fronto-parietal networks). Additional behavioral evidence shows that components of attentional control (alerting, orienting, and filtering) relate t components of WM (capacity, precision, and interference). This proposal starts from the premise that WM arises from the contribution of aspects of attention control. Additionally, the proposal draws from a recent, novel observation that there are striking individual differences in the spectral pattern of delay-period activity (SPDPA) derived from electroencephalograms (EEG) when subjects perform a WM task. Furthermore, these differences are stable and predict individual differences in other physiological measures. The goal of this research and training plan is to investigate whether differences in delay period activity might provide insight into the neural underpinnings of individual differences in WM performance. The outlined goal involves two specific aims: 1) to test the hypothesis that individual differences in the SPDPA will reflect variability in underlying patterns of effective connectivity between critical nodes in the attentio control network, and that this variability will explain individual differences in theoretically defned components of WM; and 2) to test the hypothesis that delay period patterns of effective connectivity and cortical excitability will be sensitive to cognitive training. For Aim 1, the Applicant will record the EEG from a sample of subjects performing variants of a spatial delayed- recognition task that tests the 3 WM components, with and without concurrent single pulses of transcranial magnetic stimulation (TMS). A canonical correlation analysis will be used to relate physiological measures of delay period activity (Granger causality and TMS-based excitability) to psychometric measures describing factors of attention and WM. Aim 2 tests whether theoretically discrete components of WM may be differentially sensitive to training, and that this will be manifested in training-related modulation of the physiological correlates of working memory performance. Aim 1 data will serve as baseline measures that will be compared to retest data acquired after subjects have gone through 5 weeks of intensive WM training. Identifying how training modulates cortical connectivity in correlation with psychometric measures will give insight into the neural bases of WM. Broad translational implications include improving understanding and diagnosis of conditions involving WM dysfunction such as attention deficit hyperactivity disorder, schizophrenia, and aging. Additionally, elucidating the mechanisms supporting training effects may further its value as a therapeutic intervention and perhaps enable a priori prediction of amenability to training and individualized treatment.